ydc-claude-agent-sdk-integration

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). This skill configures and uses You.com's HTTP MCP tools (mcp__ydc__you_search and mcp__ydc__you_contents) to fetch raw content from arbitrary public websites (see SKILL.md and the templates/assets such as assets/path-a-basic.ts and assets/path_a_basic.py), which the agent injects into Claude's context and thus could carry indirect prompt-injection instructions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly configures and calls the You.com HTTP MCP endpoint (https://api.you.com/mcp) at runtime to fetch arbitrary web page content via mcp__ydc__you_search and mcp__ydc__you_contents, and that fetched content is injected into Claude's context and can therefore directly control prompts (prompt-injection) during execution.